A fourth order PDE based fuzzy c- means approach for segmentation of microscopic biopsy images in presence of Poisson noise for cancer detection.

摘要: Abstract Background and objective For cancer detection from microscopic biopsy images, image segmentation step used for of cells nuclei play an important role. Accuracy approach dominate the final results. Also images have intrinsic Poisson noise if it is present in results may not be accurate. The to propose efficient fuzzy c-means based which can also handle during process itself i.e. removal combined one step. Methods To address above issues, this paper a fourth order partial differential equation (FPDE) nonlinear filter adapted with method proposed. This capable effectively handling problem blocky artifacts while achieving good tradeoff between removals edge preservation cells. Results proposed tested on breast data set region interest (ROI) segmented ground truth images. contains 31 benign 27 malignant size 896 × 768. selected all 58 are available set. Finally, result obtained compared popular algorithms; c-means, color k-means, texture segmentation, total variation approaches. Conclusions experimental shows that providing better terms various performance measures such as Jaccard coefficient, dice index, Tanimoto area under curve, accuracy, true positive rate, negative false random global consistency error, variance information other approaches detection.